2710 mpu lockup on start. Help!!


Hello good ol' Tek guys.

You will please pardon me since this is my first attempt at obtaining help on a group, but here goes:

I'm a 70yr old two-way bench tech, started in the tube/vibrator/dynamotor PS era; Motorola and GE. Component level repair. But! all analog. (hey, I was pretty futzed when transistors started to infiltrate!) I'm very weak in digital systems.

For fun I bought a 2710 SA for repair. The seller claimed FP switches were responsive, just no display. Tube problem! ?
When I got it it was as adv, very clean. On power up, fan running, LEDs on, buttons seemed to do something. So I looked at the crt, and found damaged pins and connector on the CRT, repaired, got flash on power up, and a stationary dot! Cool! now I needed to trub. the sweep ckts, found someone had removed the U911 switch that alternates log/dig info to the display. Found replacement on Mouser, installed, got sweep! Spectrumish, but no digital stuff: no menus, measurements, couldn't tell much about where I was in system. Cool again, now I'm squirming in my chair with delight. Fed the box a 25MHZ sig and got something like a freq domain display , but all kinds of messed up. front panel did change things.
Time to look at the PS voltages/ripple. Found and set voltages (10, -11, 5, -115) and VIOLA! she popped to life! The dig portion of the display was alive, The FP was responsive, I could set things like Start and stop freq span, the cursor dot moved via the freq/marker knob OHBoy! I'm a freekin genius! Life!

The first LO YIG (didn't have a clue at all about that little ball until now) started talking at that same time: a kind of irritating mid-range buzz. I don't know if that's normal, totally ignorant about these devices, and nowhere on the net can I find any reference to YIG noise (acoustic not electrical), so let me know about that, please.

So when I was adjusting the marker freq with the knob, I got slapped: the instrument reset! From where I had adjusted it via menu, it reset to default: to full span, etc, cold start kind of setup. Any time I moved the marker VERY FAST it would do a "hot-start" kind of reset back to default. By hot start I mean no time elapsed between my setup, turning the knob, the message flash and the default screen. It would be more patient if I moved the cursor slowly, but eventually would reset.
Sometimes it said something like "Frequency feedback loop failure" and show nvram stuff for a brief flash, other times it would just flash "NVRam installed, blah, blah" (maybe rev. no.s, too fast to read), and reset as above.
BUT! I had V/H sweep, frequency related functions, and could see my signal. Even had the digital stuff. So enough for that day.

Next day I powered up and checked PS voltages (all good), none seemed to change much on reset problem. So back in it's case, I let it run for a couple of hours to thoroughly warm up, and when I came back to the bench, there was no YIG noise, no display, and a completely non-responsive FP. &^%$#@!. When I restart an interesting thing occurs: brief flash on CRT, and a click around the YIG, and a few LEDs close to the display light up. When I repeat power on, the LEDs that are lit increase, from the display out to the left. A few are added each time, until all are lit. There is a shift register involved that talks to the FP for these LEDs, so...? But the FP is non-responsive. There is a second click around the YIG after a few seconds. That's it.

Opened it up, checked PS Vs, no change. With an oscilloscope I looked at the MPU lines at the edge connector on the MPU board. No digital action whatsoever. I monitored address lines and restarted the machine, and do see brief pulses of data on start, then dead. Almost all data lines are either low, or high impedance. no action on any other pin. VDD is good (4.78 V), good grounds. Seems like the MPU is ok, but something is causing it to halt operation.
There are lines, A9, A12, , A13, A18 which are high. As for the control signals, Low RD, low hold, DT/R and low InitA are high.
Haven't checked the clock yet.

I am way out of my depth here, but it seems something is sure stopping the MPU; a bad checksum somewhere? I pulled the ROM and read them (and saved images JIC!)..meanwhile I checked continuity of the bus between ROMs and checked the 40? condutor cable that is the bus between boards. BTW, battery backup (nvram) = 3V, and there is a tiny bit of corrosion product showing up as a white powder, but no trace damage (yet!)...if I get this running again, Then I'll address that changeout. I'd love to replace the NVRam with true NV ram like those Ramtron chips..(I do pinball)....

I don't know what the start sequence is: yeah, the SM says that the Frammer Cumulator is sequenced in step and added to the reet, and if this total is relevant then the ....HELP! ELIFive! I.E. I can't follow the jargon, unfortunately.

Sorry for the book, but it's the history so far. Please Spectrum Analyzer Digital Gurus, I need you, Obi-Wan Kanobi!

With the greatest of respect,
Rod Davis

PS, it would be a buzz killer, but does anyone know if Tektronix is offering repair/calib. on these? They are for older DSOs I read.


My first guess is that you have gotten close to the problem; the A10 microprocessor board.
I haven't seen very many problems with the A10 board, but when they break everything acts weirdly.
Tough part is that you are dealing with someone's failed repair and who knows what they did to the unit.

If you have downloaded the manual (there are several versions, get the latest version and work backwards to your serial number) scope check the RST line on the microprocessor, supply voltage, clock signal, and bus activity. Wait for the fail and then see what signals change. If you don't want to sit there all day, video record the scope screen. Intermittent problems are the hardest to find.

Surface mount parts can also break or come partially unsoldered, so do a good visual inspection of the A10 board. Also make sure that there aren't any loose connectors, cables go to the right connectors, or mechanical issues that can cause problems.

Start here and report back.

BTW, if Tek can repair/calibrate (not likely, they don't have spares and personnel) it would probably cost way more than buying a good refurb.


On Wed, Dec 2, 2020 at 04:03 PM, <Rodric@earthlink.net> wrote:

VDD is good (4.78 V)
I don't know this instrument. If any TTL logic is fed by this "VDD" (usually a MOS supply voltage), it'd be on the low side (Vccmin=4.75V). Probably even lower further downstream.



Just a thought Rod...

I don't know where you are located but if someone in your area has a 271x you could borrow the A10 microprocessor board from, you could try subbing it to see if that corrects or impacts your 2710's problem(s). If you do swap boards, make sure you also use your original EPROMS in the substitute board since different EPROMS may not match options in your 2710.


On Wed, Dec 2, 2020 at 08:27 PM, Raymond Domp Frank wrote:

If any TTL logic is fed by this "VDD" (usually a MOS supply voltage), it'd be
on the low side (Vccmin=4.75V).
... and not just TTL, I should have said. Things like EPROMs also have a +/- 5% voltage spec.


Stephen Bell

A failing EPROM can be very sensitive to the supply voltage being a bit low.

You can check this with one of the better EPROM programmers that has voltage marginning to verify the data. Read in the data at nominal 5V Vcc and re-verify the data at Vcc +/- 10%


On Wed, Dec 2, 2020 at 11:53 PM, Stephen Bell wrote:

A failing EPROM can be very sensitive to the supply voltage being a bit low.

You can check this with one of the better EPROM programmers that has voltage
marginning to verify the data. Read in the data at nominal 5V Vcc and
re-verify the data at Vcc +/- 10%
Many (early) devices have a Vcc spec of 5V +/- 5% (like standard TTL, Intel 2732 etc.), so 4.78V is marginal in any case. Also, It's low for a +5V nominal power supply.
I'd really check that.